Automatic analysis of carpal angles using dynamic CT imaging: Reference values in healthy wrists and assessment of scapholunate ligament injuries.

INTRODUCTION

Dynamic CT imaging is a promising modality for evaluating wrist pathologies like scapholunate ligament (SL) injuries. The primary objective of this study is to extract carpal angles from dynamic CT datasets using an automated motion analysis algorithm to provide reference values for healthy wrist motion. Secondly, the feasibility of this automatic method to detect SL ligament pathology was evaluated.

METHODS

Dynamic CT scans of healthy wrists and wrists with arthroscopically-confirmed complete SL injuries (Geissler IV) were analysed. Each scan consisted of one static image and two dynamic imaging sequences: wrist radial-ulnar deviation (RUD) and flexion-extension (FE). Bones were automatically segmented, and the radioscaphoid (RSA), scapholunate (SLA), capitolunate (CLA), and radiolunate (RLA) angles were automatically determined in each wrist position. A linear mixed model was applied to compare carpal angles between the two groups (p < 0.05).

RESULTS

A total of 84 wrists scans were analysed, of which 73 healthy and 11 injured. Reference values for healthy wrists were provided, with an average and maximum 95% CI width during all movements of 5°and 7°, respectively. Feasibility analysis showed that the SLA, CLA, and RLA were different between the healthy and injured groups during all movements. No differences were found for the RSA.

CONCLUSION

Reference values of the moving wrist of healthy participants were automatically extracted. Furthermore, our results suggest that the RLA, CLA, and SLA may be useful parameters for distinguishing wrists with complete SL injuries from healthy ones, making the automatic approach feasible.